디버그용 페이지 스크린샷 (PNG).
AI agents call web_screenshot to retrieve information from K-Personal MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Screenshot capture is fundamentally a read operation—it passively retrieves visual information from the display without creating side effects, modifying data, executing code, deleting resources, or moving money. Even though screenshots can theoretically reveal sensitive information on screen, the tool's categorization depends on what it does (capture), not downstream risks.
From the tool's definition Tool name 'web_screenshot' and description indicate it captures a screenshot of a webpage in PNG format for debugging purposes. The description uses Korean '디버그용 페이지 스크린샷 (PNG)' which translates to 'page screenshot for debugging (PNG)'.
Attacks that exploit this kind of access
디버그용 페이지 스크린샷 (PNG). It is categorised as a Read tool in the K-Personal MCP MCP Server, which means it retrieves data without modifying state.
Register the K-Personal MCP server in PolicyLayer and add a rule for web_screenshot: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches K-Personal MCP. Nothing to install.
web_screenshot is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the web_screenshot rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for web_screenshot. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
web_screenshot is provided by the K-Personal MCP server (lee30934-byte/k-personal-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →